OTp20 riak3 meeting agenda

agenda

Some possible agenda points:

- Starting point for work (ted branchen? Basho develop? 2.2???)
- existing work ( Ted got about 80% through this work once, and there exists riak_core_ng)
- backwards compatibility
- assigning work, processes and coordination (many hands make light work, but too many cooks etc)
- testing and coverage and confidence (are we sure this change has no semantic impact? Is this code even tested?)
- Riak_test or CT or wat? (related to above)
- Etc?

Random ideas:

1. Triage bugs/outstanding pull requests.
2. Smallest feature set release possible.
3. Test framework should be easy to understand and get started on - perhaps property testing could be separated from making a release - Russell mentioned getting access to Quviq open source integration server? There must be some traditional integration testing that can be done without having the license? Would like to get some hardware running and performing these tasks.
4. As per Russell's point above - it would be nice to get Heinz riak_core changes merged in - perhaps this should be top priority - what is the canonical fork of Riak right now?
5. Perhaps aim to get the experimental stuff in the following build (2.3.6?).

Future:

1. Copy gen_fsm to a dependent repository - compile and add to path if running on OTP > 19.
2. Start adding in experimental features one per build (2.3.7, 2.3.8 .. ).

Thought this columnar dep graph might be useful. Starting from the edge nodes (0) and working in. Made sense in my crazy mind when thinking / planning.

0	1	2	3	4	5	6	7	8
basho_stats	folsom	cuttlefish	bitcask	riak_core	riak_api	riak_kv	riak_repl	riak
bear	hyper	exometer_core	clique		riak_control		riak_search
canola	lager	merge_index	riak_repl_pb_api		riak_pipe		yokozuna
cluster_info	lager_syslog	riak_ensemble	riak_sysmon
ebloom	parse_trans	riak_pb	riakc
edown	protobuffs	riaknostic
eleveldb	riak_auth_mods
eper	setup
erlang_js	sext
erlydtl	webmachine
eunit_formatters
fuse
getopt
goldrush
hamcrest
ibrowse
kvc
meck
mochiweb
neotama
node_package
pbkdf2
poolboy
proper
ranch
rebar_lock_deps_plugin
riak_dt
sidejob
stdlib2
syslog

nice @martincox, it's a clear indication of where to start/where work is not going to be in conflict

@russelldb yup, makes it a lot more clear in my mind.

nice idea, good use of graph theory

Below is the list of repos with "rebar3" branches from my notes, with version notations, as of maybe last March-ish.
There are a couple more that are essentially done but not on a top-level "rebar3" branch.
I've only just started fiddling with the code again, and I've never run it through the unit tests on OTP-20 GA (though I don't expect any problems there).

%   {cuttlefish,        {branch, "rebar3"}},    % √ 16-20
%   {basho_stats,       {branch, "rebar3"}},    % √ 16-20
%   {bear,              {branch, "rebar3"}},    % √ 16-20
%   {clique,            {branch, "rebar3"}},    % √ 16-20
%   {cluster_info,      {branch, "rebar3"}},    % √ 16-20
%   {edown,             {branch, "rebar3"}},    % √ 16-20
%   {eleveldb,          {branch, "rebar3"}},    % √ 16-20
%   {exometer_core,     {branch, "rebar3"}},    % √ 16-20
%   {folsom,            {branch, "rebar3"}},    % √ 16-18   QC fudged, OTP-19+ eunit crash
%   {getopt,            {branch, "rebar3"}},    % √ 16-20
%   {goldrush,          {branch, "rebar3"}},    % √ 16-20
%   {hamcrest,          {branch, "rebar3"}},    % √ 16-20
%   {lager,             {branch, "rebar3"}},    % √ 16-19   OTP-20 errors
%   {meck,              {branch, "rebar3"}},    % √ 16-20
%   {neotoma,           {branch, "rebar3"}},    % √ 16-20
%   {parse_trans,       {branch, "rebar3"}},    % √ 16-20
%   {pbkdf2,            {branch, "rebar3"}},    % √ 16-20
%   {rand_compat,       {branch, "rebar3"}},    % not using
%   {riak_core,         {branch, "rebar3"}},    % √ 16-20
%   {riak_ensemble,     {branch, "rebar3"}},    % √ 16-20
%   {riak_sysmon,       {branch, "rebar3"}},    % √ 16-20
%   {setup,             {branch, "rebar3"}},    % √ 16-20
%   {chronos,           {branch, "rebar3"}},    % √ 16-20
%   {hyper,             {branch, "basho"}},     % √ 16-20
%   {sext,              {branch, "rebar3"}},    % √ 16-20
%   {canola,            {branch, "rebar3"}},    % √ 16-20

btw, I've started twiddling here: https://github.com/rebar3riak

@tburghart I'd heard about these, like in @russelldb's agenda points, but had no idea there were so many. I'd just looked in branches for "otp", never noticed the rebar3 ones (doh!). Awesome starting point 👍

except what branch are they based on? develop? 2.2?

not sure if 1 or 2 pre-releases/release-candidates of 3.0 (3.0-rc-1 & 3.0-rc-2 perhaps?) with the number of possible breaking changes, in performance as well, which users can try out before the official launch of 3.0?

notes:
1 . take what Ted has done as a starting point
2. what Ted hasn't done work on riak-2.2 repos
3. do not have backwards compat work going forward
4. martin cox to add simple work coord page to riak repo
5. porting riak test to CT whatever is another battle
6. meet 1st wed of the month next month for progress